Article

Journal of Research in Crime and

Delinquency
2018, Vol. 55(1) 3-26
Weapons, Body ª The Author(s) 2018
Reprints and permission:
Postures, and the sagepub.com/journalsPermissions.nav
DOI: 10.1177/0022427817706525

Quest for Dominance journals.sagepub.com/home/jrc

in Robberies:
A Qualitative Analysis
of Video Footage
Floris Mosselman1, Don Weenink1,
and Marie Rosenkrantz Lindegaard2

Abstract
Objective: A small-scale exploration of the use of video analysis to study
robberies. We analyze the use of weapons as part of the body posturing of
robbers as they attempt to attain dominance. Methods: Qualitative analyses
of video footage of 23 shop robberies. We used Observer XT software
(version 12) for fine-grained multimodal coding, capturing diverse bodily
behavior by various actors simultaneously. We also constructed story lines
to understand the robberies as hermeneutic whole cases. Results: Robbers
attain dominance by using weapons that afford aggrandizing posturing and
forward movements. Guns rather than knives seemed to fit more easily
with such posturing. Also, victims were more likely to show minimizing

1
Universiteit van Amsterdam, Amsterdam, the Netherlands
2
Netherlands Institute for the Study of Crime and Law Enforcement, Amsterdam, the
Netherlands

Corresponding Author:
Don Weenink, Universiteit van Amsterdam, PO Box 15508, 1001 NA Amsterdam, the
Netherlands.
Email: d.weenink@uva.nl
4 Journal of Research in Crime and Delinquency 55(1)

postures when confronted with guns. Thus, guns, as part of aggrandizing

posturing, offer more support to robbers’ claims to dominance in addition
to their more lethal power. In the cases where resistance occurred, rob-
bers either expressed insecure body movements or minimizing postures
and related weapon usage or they failed to impose a robbery frame as the
victims did not seem to comprehend the situation initially. Conclusions:
Video analysis opens up a new perspective of how violent crime unfolds
as sequences of bodily movements. We provide methodological recom-
mendations and suggest a larger scale comparative project.

Keywords
robberies, violent crime, interactionism, video analysis

There is now a tradition of research into robberies of various forms (in

private homes, shops, or on the street), which seeks to understand how these
interactions unfold. A classic is Luckenbill’s (1981) study of how shop
robbers try to frame the situation as a robbery. This work shows that
whether robberies end up in violence or not depends, among others, on how
successful robbers are at imposing a “working agreement” on the victims:
They should not just accept that they are victims, but they should also
understand and fulfill the role that robbers have in mind for them in various
stages of the robbery situation. Subsequent work has elaborated on these
stages (Bernasco, Lindegaard, and Jacques 2013; Copes et al. 2012; Jacobs
2012), and crucially, the use of weapons in relation to the degree of violence
involved (Cook 1981; Feeney 1986; Kleck 2005; Kleck and Delone 1993;
Kleck and McElrath 1991; Wells and Horney 2002). In this article, we delve
further into the robbery situation. Based on video footage of shop robberies,
we provide close-up analyses of the bodily and emotional dynamics of
robberies and the role of weapons therein. While our study is explorative,
also given the small sample size, the contribution we aim to make is
threefold.
First, while prior analyses have relied on interviews and judicial case
files to reconstruct robbery situations, we use real-time visual data. This
material not only allows for the validation of the results of prior studies that
were based on these other data sources but it also enables us to probe deeper
into the robbery situation by repeated and variable viewing (slow and fast
motion, forward and backward) of the event.
Second, previous interactionist work did not consider the role that bodies
and emotions play in robbery situations. To our knowledge, only
Mosselman et al. 5

Lindegaard and colleagues (2013) paid attention to emotions in robberies.

However, their study did not focus on how emotions impact the dynamics of
the situation. Our assumption is that emotions play an important role in the
establishment of a working agreement between robbers and victims and
hence influence the outcome of the robbery. We aim to infer (changes in)
the emotional states of robbers and victims from the sequences of body
movements and postures they display, specifically with regard to the use of
weapons.
Third, with respect to weapons, prior U.S. work has demonstrated that
the display of (lethal) weapons by robbers tends to reduce the severity and
likelihood of violence (Cook 1981;Kleck and Delone 1993; Luckenbill
1981; Wright and Decker 1997; but see Bernasco et al. 2013, who report
that in Dutch robberies, weapon use by robbers increases the likelihood of
violence). We intend to specify this work by investigating how exactly
weapons are being used. We are going to argue that in robbery situations,
weapons should be seen as body extensions, which are crucial in the rob-
bers’ attempts to attain “emotional dominance” (Collins 2008) which
enables them to impose a robbery frame on the situation. Emotional dom-
inance is the sense that one can impose one’s will unto the other, that one is
able to make the other follow one’s rhythm, and that the other will not resist
(any longer).
The main question we aim to answer is how does the use of weapons as
part of robbers’ body posturing influence their attempts to attain emotional
dominance? Our study is empirically grounded on Closed Circuit Televi-
sion (CCTV) footage of 23 armed robberies available at the Netherlands
Institute for the Study of Crime and Law Enforcement. In what follows, we
will first outline the conceptual framework and specify the research ques-
tions. After that, the data, sampling, and procedures of data analysis are
explained. Two sections report on the empirical results, each focusing on
two distinctive stages in robberies. In the concluding section, we discuss the
use of video data for the analysis of violent crime in general and robberies in
particular.

The Interactional Dynamics of Robberies and the Role of

Weapons Therein
Seen from the interactionist perspective on crime (see Felson 1982; Felson
and Steadman 1983; Felson and Tedeschi 1993, Wilkinson and Fagan
2001), robberies, like all interactions, are performances in which actors try
to present a situational identity; a claim on how others should value and
6 Journal of Research in Crime and Delinquency 55(1)

treat them (Goffman 1959). Thus, a successful framing of the situation

(from the viewpoint of the robbers) requires cooperation between robbers
and victims, while avoiding open antagonism or resistance. The task for
robbers is to impose a “working agreement” (Polk 1999) on the victims
which defines how they should interact in the robbery situation. As noted
above, this requires emotional dominance on the side of the robbers (Collins
2008).
Analyzing 261 robberies, Luckenbill (1981) identified four stages that
robbers and victims must accomplish together in order to create a successful
robbery, that is, a robbery in which the robbers get what they want at the
least possible costs in terms of the use of threat, violence, and antagonistic
tension—assuming that shop robbers follow instrumental reasoning, which
may not always be the case. First, the robbers must move into copresence
with the victims; second, the victims and robbers must transform their
encounter into a robbery frame; third, one or both parties must transfer the
valuables; and fourth, the offenders leave the setting with the valuables.
This article focuses on the first two stages only for two reasons. First, if
robberies fail, this occurs most frequently in the second stage (cf. Lucken-
bill 1981), making this stage the most important for the outcome of the
robbery. Second, the second stage appears to be the longest, the most
violent, and complex in most robberies and therefore seems most relevant
to our purpose.
Let us take a closer look at Luckenbill’s first two stages of the robbery
situation. The first stage, moving into copresence with the victims, means
that the robbers move into striking range without raising suspicion to avoid
perhaps unmanageable opposition (Luckenbill 1981:29). Robbers tend to
use two tactics in this stage (see also Wright and Decker 1997:100). One of
them is “speed and stealth”: The offenders prepare for the strike without
being noticed by the victims (the stealth aspect) and then suddenly they
move into copresence with the victims, having their weapons ready (the
speed aspect), “disguise” is the second tactic. It involves the hiding of the
robbers’ intentions, so that the victims perceive them to be a “legitimate
part of the setting” as customers or as passersby who request information
(Luckenbill 1981:29; Wright and Decker 1997:100).
In the second stage, the robbers must transform the situation into a
robbery, which involves “a succession of moves between the offender and
victim” (Luckenbill 1981:31). First, the victims should suppress opposition,
and the offenders must be in control of the behavior of the victims by means
of force or the threat thereof, they must be emotionally dominant. One way
to impose the robbery frame is by using incapacitating force or what
Mosselman et al. 7

Luckenbill described as “bodily pain which debilitates or immobilizes the

victim for a time” (1981:31). Another, more common way is to express a
command, backed up by a threat. Luckenbill also found that the more lethal
the weapons used by the offenders, the more likely the victims submitted.
This is also his explanation for why so many offenders with nonlethal
weapons like clubs or bare hands open with incapacitating force. They do
this to gain dominance despite their lack of lethal punitive resources.
Later quantitative work in the United States confirms these findings.
Kleck and Delone’s (1993) study on weapon use in robberies showed that
the more lethal the weapons displayed, the less opposition and violence
occurred. They also found that the likelihood of victim injury was associ-
ated with victim resistance by means of physical force or weapons other
than guns, but not when victims used guns (Kleck and Delone 1993). So
when either offenders or victims wield a gun, violence and injuries are
reduced (see also Wells and Horney 2002). However, these findings were
not confirmed by Bernasco et al.’s (2013) study of 256 robbery situations in
the Netherlands based on interviews with 100 robbery offenders. They
found that weapon use increased the risk of violence during robberies.
Bernasco et al. (2013) showed that in 30 percent of the robberies, the
offender started the robbery by using violence with the use of weapons,
while in 70 percent of the cases, the offender used violence at a later stage
only. While violence at the start of the robbery was not related to victim
resistance, the violence in later stages occurred mostly (50 percent of the
cases) after the victims resisted as a means of (re)enforcing compliance
(Bernasco et al. 2013).
While Bernasco et al.’s (2013) study of robberies in the Netherlands
focused on whether a weapon was exposed during the robbery and in what
stage it was exposed, studies of weapons and robberies in the United States
focused on the effect of the presence of weapons in robberies (Cook 1981;
Kleck 2005; Kleck and McElrath 1991; Wells and Horney 2002) without
specifying how weapons were used in the course of the event. For the
understanding of robberies, it seems crucial to know how weapons are
brought into play in the interactional dynamics of robberies, especially in
the United States, where carrying a weapon is more common.
From a phenomenological approach, robberies are a form of intentional
navigating in the world; robbers have projected a trajectory which they
carry out by moving their bodies. We argue that weapons play an important
role in the robbers’ embodied intentionality. Not only because weapons are
charged with meanings that support the robbery performance but also
because they literally “point” the bodies of victims and those of the robbers
8 Journal of Research in Crime and Delinquency 55(1)

themselves toward a specific set of actions. In Heideggerian (1927/2002)

terms, the weapon must change from present-at-hand to ready-to-hand. This
means that weapons can only support or direct robbers’ intended maneu-
vering to the extent they are experienced as body extensions, as integral to
the robbers’ embodied being. Thus, we understand the carrying, displaying,
and actual use of weapons as part of sequences of body movements and
postures. This also means that weapons can both support and impede rob-
bers’ intended course of action. Let us conclude with a note on the stand-
point from which we analyze the data. Both Luckenbill’s conceptualization
of the stages and tactics in robberies and Collins’s theory about the emo-
tional dynamics of violence are external to the actors studied. While this
study is thus conceptually grounded on a “data inspectors’” standpoint, we
combine it with a phenomenological or “participants’” standpoint—in this
case that of the robbers’.1 More specifically, we are going to analyze how
robbers carry, display, and use weapons as body extensions in their attempts
to establish emotional dominance and in various stages of the robbery.

Method
Data and Sample
The data are CCTV video footage of shop robberies in the Dutch cities of
Amsterdam and Rotterdam. The choice for these two cities was based on
practical reasons, as more than one third of all reported robberies in the
Netherlands are committed in these two cities (Overvallen NL 2016).
Access to the material was provided by the Dutch Ministry of Justice and
the Rotterdam and Amsterdam police forces. All robberies in the sample
were therefore reported to and investigated by the police. The analyses were
conducted under strict judicial conditions considering the privacy of all
visible persons in the video. For this reason, we had to transform the video
stills included in this article into a cartoonlike quality. Unfortunately, we do
not have permission to provide a link to the raw data. The data files were
stored on encrypted hard disks that were kept in a safe.
The material was collected by the last author (see also Lindegaard et al.
2016). We received footage of 127 Rotterdam cases and 48 Amsterdam
cases committed in 2013 and 2014. The robbery team of the Rotterdam
police force uses a central computer, on which all their CCTV footage is
stored. We received a copy of their data as stored in August 2014. In
Amsterdam, robbery footage is dispersed among various criminal investi-
gations teams. As the collection of these data would require too much of our
Mosselman et al. 9

time and energies, we turned to the External Communications Department

of the Amsterdam police force. This department provides and edits video
material for communication to the public, asking for their assistance in
investigations, among others via a national television program (Opsporing
Verzocht). In this study, we used the original and unedited footage.
Of the 175 cases we received, 47 had to be dropped because the video
was not working or missing, and in 78 cases, the quality of the video was
poor, or there was no or only part of the interaction visible. In the latter
cases, we typically found recordings of cars passing by outside the shop, or
the footage showed the offender entering the shop, but it did not show what
happened inside the shop. Eventually, only 50 cases remained that were
suitable for analysis.
Apart from Collins’s (2008) analysis of violent interactions based on
photo material and Klusemann’s (2010) video analysis of the interaction
between Dutch UN military leader Karremans and his Serbian counterpart
Mladic preceding the Srebrenica war atrocity, no qualitative studies of
crime based on video data had been published at the moment we conducted
this research, so that we could not rely on established methodological
practices or exemplars. Consequently, we had to explore various analytical
strategies, and some of them turned out to be less useful than we had hoped.
As a result, due to time constraints, this article is based on 23 randomly
selected cases of the remaining 50. However, as our analysis will show,
these 23 cases were a sufficient number to analyze how weapons use and
sequences of bodily movements influenced the robbery situation.
It should be noted that we cannot assess to what extent our sample is
representative for the approximately 2,000 to 3,000 robberies which are
committed in the Netherlands yearly (Rovers et al. 2010). Our data might
be biased in several respects. First, not all shops that are targeted by robbers
have CCTV cameras installed. Second, the Amsterdam cases, for which
assistance of the public was sought, might contain more severe robberies in
terms of violence and social impact. However, we do not consider these
potential selection biases highly problematic, as our aim is not to arrive at
results that are representative for Dutch robberies in general.
To get a sense of the sample, we now provide descriptive statistics on
key variables such as the types of weapons carried by the robbers, the
number of robbers and victims, the time duration of the robberies, and the
locations were the robberies took place. Table 1 displays the types of
weapons offenders and co-offenders carried.
The sample provides sufficient variation with regard to weapons, given
our aim to inquire how they may affect robbers’ body posturing differently
10 Journal of Research in Crime and Delinquency 55(1)

Table 1. Types of Weapons, Carried by Robbers.

Weapons Number

Guns 13
Knives 10
Hammers 4
Other/not clear 3

Table 2. Types of Locations, Targeted by Robbers.

Locations Number

Night shop, small supermarket 4

Jewelry store 4
Large supermarket 4
Hotel 4
Other 7

in their attempts to attain emotional dominance. While 13 offenders and

co-offenders carried guns, 10 displayed knives and 4 used hammers (in one
case, a hammer was used to threaten a cashier only; in another case, three
hammers were used both to smash display shelves and to threaten the shop
owners). In two cases, the weapon was not clearly visible, and one robber
used a chair as a weapon.
The average number of offenders was 1.7, and the maximum was 3.
Similarly, the average number of victims was 1.6, the maximum being 3.
The robberies lasted between 10 seconds and 20 minutes with approxi-
mately one third lasting less than a minute, one third lasting between 2 and
3 minutes, and the remaining ones more than 3 minutes. Table 2 provides an
overview of the locations where the robberies took place.
It can be inferred from Table 2 that the robbers in our sample targeted a
diverse range of locations. The seven other types of businesses were a
telecom shop, flower store, taxi, café, cinema, a clothing store, and a gas
station.

Analytical procedures and analytical problems. Video analysis allows research-

ers to map different modes of bodily behavior by various actors simultane-
ously, thus providing an appropriate tool for the close-up analysis of social
interaction (Cowan 2014:7; Knoblauch 2009). Moreover, the opportunity to
Mosselman et al. 11

Figure 1. Visualization of coding with Observer of robbery situation no. 1.

repeatedly view the data at various speeds (ranging from very fast to very
slow or stills) and in various directions (backward, forward) is of great
value; this enables to verify the coding, thus increasing data validity
(Luckmann 2006). In our research, it occurred that in some cases, even
after carefully observing the footage in slow motion, new discoveries were
made as behaviors were overlooked first. This is because video data are
very complex, especially when it concerns multiple actors who engage in
different modes of behavior. The focus of analysis in our case was on
sequences of body movements that transform the situation; every move-
ment indicating any emotional dynamic (based on literature) was included
in the coding (see Ekman 1991; Hall, Coats, and Lebeau 2005; Kluseman
2010; Scheff and Retzinger 1991). Body movements that were interrupted
because actors moved out of reach of the camera were not included in the
analysis.
We created a codebook that contained codes for body postures and
movements that indicated either emotional dominance or submissiveness
(see Appendix). Our codebook was based on preliminary analyses of the
same footage with other (student) observers and prior research on the emo-
tional meanings of bodily behavior (Collins 2008; Hall et al. 2005; Kluse-
mann 2010; Scheff and Rentzinger 1991).
We used Observer XT software to code body postures and movements.
This program allows to code different modes of behavior of different indi-
viduals at different points in time (see Cowan 2014:15). Figure 1 provides a
visualization of the coding of one, relatively simple, robbery situation in
Observer.
Figure 1 shows two robbers who first display minimizing postures. At
point 12, robber 1 moves forward to victim 1 and produces a weapon. The
victim responds by moving away. At that moment, robber 1 threats the
12 Journal of Research in Crime and Delinquency 55(1)

Two offenders enter the shop with their hoods on while the victim is in the
store, offender 1 leading the way. They wait in front of the cash register as if
they want to order something. Both actors are looking down [both robbers
minimize posture]. Offender 1 has his left hand in his pocket. The victim
moves behind the cash register, looking at the offenders. Offender 1 puts
some money on the counter, still controlling his posture, looking down, not
making eye contact. Offender 2 stands behind offender 1. The victim points
out a candy to take for offender 1. When offender 1 takes the candy offender
2 synchronously moves with him, keeping himself behind the first offender.
When offender 1 takes the candy offender 2 again synchronously moves with
him, staying behind the first offender. After putting the candy on the counter
offender 1 suddenly rushes around the counter, towards the victim while
drawing his knife [robber 1 moves forward and shows weapon]. The victim
feigns back when offender 1 moves in on her [victim 1 moves back]; knife
drawn, arm stretched straight forward, aggressively pushing her back. During
this action offender 2 is looking inside the shop, not at what is happening
[maintaining mimized body posture]

Figure 2. Story line of robbery situation no. 1.

victim. All the time, robber 2 displays a minimizing posture. The strength of
using Observer is that it ensures high levels of coding reliability and offers
tools for quantification of predefined behavioral categories. The weakness
is that, due to the fine-grained coding, the situation as a whole—which
provides the meaning context to understand body posturing and move-
ments—disappears from view. To give an example, the meaning of victims
raising their hands is not unequivocal; it could mean either to submit to a
threat, to actively counter a threat, or to make a calming gesture toward the
robber, depending on how the situation unfolds.
Rather than increasing the complexity of the coding scheme by incor-
porating many different codes for each meaning of raising hands and var-
ious other bodily movements, we decided to write “story lines” to capture
each robbery situation as a hermeneutic unit. This decision was motivated
by our research question, focusing not on the effect of weapon presence but
on how weapons and other forms of body posturing and movements influ-
ence the process of establishing dominance. The story lines depict the
detailed course of action of every actor involved, including our interpreta-
tions of these actions as they follow up on one another. Figure 2 provides an
example of the story line of the same robbery as depicted in Figure 1, which
showed the time lines created with the Observer software.
Mosselman et al. 13

Figure 3. A robber uses the speed and stealth tactic as he enters a supermarket
swiftly. He looks downward at the ground, neck and shoulders slightly bent down,
holding a hammer downward in his left hand as he rushes toward the cashier (still
taken from case 19).

While the Observer software seduces researchers to fragment video

data into predefined categories, the construction of story lines invites
researchers to come up with categories inductively, fitting to the situation
as a whole. As our story lines involved more interpretative work (e.g.,
rather than “raising hands” descriptions were now of the type
“aggressively pushing her back”), we went back and forth comparing the
story lines with the video footage to increase reliability. In cases we dis-
agreed on the interpretation, consensus was established by repeatedly
watching the footage.
We analyzed the story lines in five stages. First, all the cases were
divided into the first two stages of robbery described by Luckenbill
(1981). Second, all the relevant behaviors that signaled transformations in
emotional states were highlighted. Third, the robbery interactions were
analyzed in a vertical way based on the multimodal transcripts coded in
Observer, now focusing on the sequences of body movements between
robbers and victims. Fourth, the robberies were analyzed in a horizontal
way, analyzing the sequences of interactions per robbery stage, based on the
multimodal transcripts again. Lastly, all these material were grouped and
compared to find patterns with regard to the carrying, display, and use of
weapons, in different stages of the robbery.
14 Journal of Research in Crime and Delinquency 55(1)

Results
The First Stage: Moving into Copresence
In this section, we report how robbers work toward an encounter with their
victims. In line with Luckenbill (1981), we identified the speed and stealth
tactic and the disguise tactic as ways in which the offenders move into
copresence with their victims. In addition to describing these forms of
establishing copresence, we report on the emotional dynamics expressed
in sequences of body movements and posturing.

Speed and Stealth Tactic

First, the speed and stealth tactic was apparent in 16 of the 23 cases studied.
In 11 of these 16 cases, the robbers had their weapons drawn when entering
the scene. Of these, seven are with guns, three with knives, and one with a
blunt object (see Figure 3). These cases include both the offenders who
quickly aimed their weapons at the victims directly after entering the scene
and robbers who kept their weapons hidden until close to the victim.
Offenders generally tended to keep knives low, at waist height, while
guns were mostly kept at shoulder height when posing the initial threat. It
also seemed that robbers were more likely to keep knives hidden until close
to the victim probably because they were concerned about alarming the
victims, as proposed by Luckenbill (1981). Compared to gun threats, threats
with a knife are less compelling from a few feet away. In 10 cases, the
victims maintained their normal routines when robbers had entered the
scene mostly because they did not spot the offender yet. However, in three
of these cases, the victims clearly saw the face-covered offenders approach-
ing, but they just froze or maintained their normal routines.
In 11 of these 16 cases, the robbers wore some sort of face cover.
Covering one’s face probably plays various functions in robberies. First,
there is the obvious goal to hide one’s face from being recognized, also
given the widespread use of CCTV cameras in shops. Second, putting on a
face cover provides a symbolic–emotional transformation for the robbers
themselves, preparing them to play their role in the upcoming action. How-
ever, sometimes face covers such as hats, caps, or loosely tied shawls seem
to inhibit the successful imposing of a robbery frame at least for a moment:
Robbers clumsily need to adjust these garments to avoid their faces being
exposed to the camera. Their fumbling and tinkering around resulted in
minimizing, self-oriented body movements. While arranging, designing,
and wearing proper disguises that support rather than impede the course
Mosselman et al. 15

of the robbers’ actions has been part of the preparatory tasks of robbers for a
long time, the widespread use of CCTV makes this task even more crucial.
Weapons were used to emphasize forward and pointing arm movements
especially when robbers proposed a threat. As extensions of the arm, weap-
ons contributed to robbers’ aggrandizing postures. In doing so, they did not
just enlarge their spatial presence but also prepared to invade the body space
of the victims. At the same time, this is their claim to emotional dominance.
Hence, weapons do not achieve compliance just by being possessed or
displayed, but only as extensions of the robber’s body as it executes move-
ments and postures that are experienced (by the robber) and recognized (by
the victim) as being part of a robbery. Domineering postures seemed to be
more associated with guns than with knives. Of the 11 robbers who carried a
gun, we noted expanded body postures 11 times, while the 6 robbers
who used knives displayed aggrandizing posturing only once (see
Figures 4 and 5). Guns seem to expand the posture of the offenders more
than any other weapon. When victims become aware of the robbers’ inten-
tions, we arrive at the next stage, in which the robbers aim to enforce and
maintain the robbery frame. First, however, we will show the results
concerning the other tactic used in the first stage of the robbery.

Disguise Tactic
In the disguise tactic, robbers attempt to take victims by surprise by acting
as regular customers. We found this tactic in 7 of the 23 robberies. In five of
these cases, the robbers acted as if they were about to make a regular shop
transaction or asked victims for information. In the other two cases, the
robbers only looked around the store before initiating the robbery. Robbers
who presented themselves as regular part of the “shop setting” share body
postures that we interpreted as lacking emotional dominance. First their
gaze: In five of these seven cases, the robbers avoided eye contact with
their victims (while the averted gaze of these robbers may result from their
efforts to remain unrecognizable given the shop camera, it may still have
the effect of being perceived as not very dominant in the eyes of victims).
Second, in five of these cases, robbers kept their hands in their pockets
when strolling around or approaching the victim. In these cases, the offen-
der took out a weapon from these same pockets when they were close to the
victim. In two other cases, the disguise tactic seemed even more counter-
productive, given the aim of imposing a robbery frame later. Here, the
robbers hesitantly walked back and forth in the shop with their bodies
16 Journal of Research in Crime and Delinquency 55(1)

Figure 4. A robber poses a threat in an aggrandizing posture. The way the robber
holds the gun, with his elbow slightly bent, extending his arm both sideways and
forward, contributes to an aggrandizing posture (still taken from case 10).

stiffened, gaze averted, and hands kept in their pockets once they made
contact with the victims.
These body postures not only seemed discordant with the outward mov-
ing, expanding mode of bodily behavior that is characteristic of attempts to
attain emotional dominance, they also did not fit very well to the normal
customer–shopkeeper interaction, in which both parties generally make eye
contact and in which customers show their hands, ready to receive or give.
Nevertheless, not once a victim recognized and countered a robber before
he would present himself as such.
Although we should be prudent in making numerical claims, the results
suggest that, contrary to the speed and stealth tactic, robbers who disguised
Mosselman et al. 17

Figure 5. Typical way of holding a knife. The still shows how robbers typically use
their knives, pointing at belly height as they pose their threat, which does not
contribute much to aggrandizing posturing (still taken from case 25).

their intentions first seemed more likely to use knives rather than guns. To
be precise, of the robbers who deployed the disguise tactic, four used
knives, two guns, and one an unidentifiable object for smashing a display
case. The use of knives in the disguise tactic also means that robbers need to
move closer to the bodies of their victims to pose their threats.
Our analyses so far suggest that the disguise tactic creates more emo-
tional and body work for robbers: They have to transform their fake and
sometimes even hesitant customer postures to the compelling body move-
ments that are needed to impose the robbery frame. Weapons seem to play
an important role in this transition, but hiding them and then displaying
them in a convincing way require substantially more emotional and body
control as compared to the speed and stealth tactic. Our data also suggest
that guns and knives have different effects on the interaction and are related
to the tactics which were used. Guns, rather than knives, seem to provide
more opportunities for offenders to make aggrandizing body movements
and postures, while robbers with knives had to be in close proximity to the
victim to pose a threat convincingly.

The Second Stage: Imposing and Maintaining the Robbery Frame

After the robbery is announced, robbers have to continue imposing their
dominance throughout the remainder of the robbery. Our analyses suggest
18 Journal of Research in Crime and Delinquency 55(1)

that two scenarios exist in the offenders’ quest for dominance in this second
stage: A scenario in which the robbers gain and maintain total dominance
throughout the robbery, and one in which their dominance is contested.
These scenarios are not necessarily related to the success of the robbery
from the viewpoint of the offenders. To give an example, even when offen-
ders were able to dominate the whole interaction with a gun, it happened
that robberies failed due to the fact that the valuables were not accessible or
the robbers just left for unknown reasons. On the other hand, it also
occurred that while robbers encountered heavy resistance from the victims,
they still got the valuables in the end.

Total Dominance
Total, uncontested dominance of the robbers occurred in 12 of the 23
robberies. In 8 of these 12 cases, the interaction started with body postures
that clearly signaled dominance by the robbers, followed by body postures
that indicated submissiveness by the victims. The fragment below illustrates
a typical case.

When he is two meters away from her, he aims the gun at her face, arm
stretched. She stands behind the counter and puts her arms in front of her
chest, moving a bit backward. (case 10)

The robber stretched out his arm and aimed his gun at the face of the victim,
expanding his body posture by raising the gun. Stretching out an arm and
pointing with a finger indicates a demand, suggesting a direction for the
others’ actions (Siegel, Friedlander, and Heatherington 1992). Pointing with
a knife, gun, or hammer intensifies the force of the demand. As we noted
above, weapons should be seen as bodily extensions that support the inten-
tional maneuvering of robbers. The response of the victim after this pre-
sentation of dominance was submissive: She moved a bit back and put her
arms in front of her chest, indicating the vulnerability of the body she was
protecting, thus emphasizing her weak position. In other cases, the victims
ducked down, moved back, or raised hands, mostly with the palms toward
the robber. These are all signals of submission and forced acceptance of the
robbery frame. To maintain the robbery frame, robbers had to affirm their
emotional dominance during the interaction, and they subsequently
expected victims to express submissiveness throughout the robbery. Con-
sequently, some robbers use their weapons to threaten victims multiple
times during the interaction.
Mosselman et al. 19

The following fragment describes how a robbery frame is kept in place

with repeated use of body movements by the robber:

Robber 1 walks to a cash register and hits a metal part of it twice. Immedi-
ately victim 1 looks at him and holds her head a bit back. She is sitting at the
cash desk next to the one the robber had hit, approximately three meters
away. The robber stands upright, looks straight at victim 1 and points his
hammer at her face, his arm stretched. He then turns his back to her and walks
a small circle, looking at robber 3 who came in behind him. Victim 1 feigns
back a bit further. . . . Victim 1 slowly gets out of her chair and walks away.
Now robber 1 looks at her and moves in her direction, again pointing his
weapon. Then he shortly points at the cash register in the back, after which he
aims it at her again. He now stands against the cash register, pointing his
hammer at the chair in front of it. Victim 1 moves back to her chair and robber
1 stops aiming his hammer and fumbles to open a bag. (case 19)

Initially, robber 1 established dominance by using his weapon on an object,

thus announcing he has a dangerous weapon. The victim responded with a
body posture that is retreating as she pulled her head back. From that point,
the robbery frame was established. But as soon as robber 1 turned his back
to her and reoriented his focus of attention elsewhere, the victim took the
opportunity to move away, trying to escape. The robber however com-
manded her back by moving forward and pointing his weapon.
In 9 of the 16 cases in which robbers used the speed and stealth tactic,
they were able to maintain dominance throughout the whole robbery using
mostly guns (six) but also knives (three) and in one case a hammer. In three
of the seven disguise tactic cases, the offenders did not encounter any
resistance and enforced and maintained the robbery frame successfully in
the second stage by using a single threat with a knife, a gun, or an unknown
object.

Contested Dominance
In 11 of the 23 cases, victims contested the robbers’ attempts to impose a
robbery frame. We identified two types of situations. First, in five cases, the
robbers did not succeed in attaining emotional dominance initially as
expressed by their own body postures and/or weapon usage. Second, in five
cases, the robbers failed to impose a robbery frame altogether—they did not
make clear to the victims what was happening: Either robbers did not give
sufficient time to the puzzled victims or they went straight for the valuables
20 Journal of Research in Crime and Delinquency 55(1)

without confronting the victims. In these latter situations, victims did not
show signs of submissiveness because they did not seem to comprehend the
situation. The following fragment shows one of the five cases in which a
victim initially resists the robbers’ attempts to impose a robbery frame:

Suddenly the robber draws a small knife, holds it in front of his belly and
moves closer to the victim looking straight at her, his head forward. The
counter is between them. The victim stays put, holds her ground, and reaches
out with one hand for a split second, countering the knife. Now the robber
repositions the knife at head height and leans forward again which makes the
victim move back a bit and raise her hands in front of her chest, palms
towards the robber. (case 4)

As in the cases where the robbers remain dominant throughout the whole
interaction, the robber in this example opens with a threat. Even though he
looks straight at the victim and bends his head forward, the way he carries
the knife—keeping it low, close to his belly—does not fit with the threat,
which requires a more expanding, aggrandizing bodily posturing. Here, the
knife does not support the intentions of the robber. Or, more precisely, by
using the knife in this way, the robber displays a certain hesitance, a lack of
unequivocal intentionality. The victim does not move back nor minimizes
her posture in any way, she even moves forward with her hand as if to
counter the knife. However, the robber manages to reclaim dominance by
repositioning his knife at head height, by expanding his posture, and by
making a move toward the victim. Now the victim acknowledges that sub-
mission is what the robber has in mind, by raising her hands in front of her
chest, palms toward the robber, and moving back.
The other cases demonstrate that imposing a robbery frame crucially
requires a working agreement between robbers and victims. Here, the
robbers seemed too eager to impose dominance without noticing that the
victims did not understand what happened to them. Consequently, they did
not show the expected signs of submissiveness. This happened in the
following example. The robber enters the scene with his gun drawn:

He points the gun at victim 1 while walking in on him fast. Victim 1 does not
seem to respond, he keeps exactly the same posture he had before robber 1
entered the scene. Together with victim 2, victim 1 stands behind the cash
desk. Also, victim 2 does not seem to react and stands with his hands behind
his back. Robber 1 now points his weapon at victim 3, who stands in the
middle of the store. Victim 1 takes a step back as if his reaction to the
Mosselman et al. 21

experienced gun threat was delayed. Robber 1 reaches victim 3, puts the gun
on the temple of the victim’s head and pushes him backwards violently with
his gun. Victim 3 initially only moves back under force, but then he grabs the
robber and resists. They start a struggle in the back of the shop. (case 8)

Note that in real time, this part of the interaction took less than two seconds.
The robber gives no time for victim 1 to respond and then immediately
refocuses on victim 3. He misses the submissive signal of victim 1 who
steps back and rapidly poses a severe threat to victim 3 by putting the gun on
his head. He also starts pushing victim 3 immediately who then resists
grabbing the robber. Consequently, the robbers hastily increase the severity
of their violent threats. In some cases, this resulted into more resistance by
victims.
As the sample size is small, we cannot rely on statistical measures to test
relationships. However, our qualitative exploration does show some ten-
dencies we want to highlight here. First, most weapon threats support the
forward moving by the offenders. Second, it seems that guns allow for
expanding body postures more easily than knives. As we have noted above,
robbers tend to hold knives at belly height when they approach their vic-
tims, while guns are more often displayed at shoulder height, requiring
more extension of the robbers’ arms and thus expanding their bodies. Thus,
apart from their more lethal power, the appropriate handling of guns
requires a larger body space, which provides additional support to robbers’
attainment of emotional dominance (see Figures 4 and 5).

Conclusion
This is one of the first small-scale and explorative studies of robberies using
video data. It shows that for a robbery to be successful from the viewpoint
of the robbers, emotional dominance is required to impose a working agree-
ment on the victim. Robbers attain emotional dominance by combining
expanding body postures and forward body movements with weapons that
afford and stimulate these movements. Robbers who use the disguise tactic
to move into copresence with the victims face the challenge to transform the
situation into an asymmetrical one as they try to impose the robbery frame.
This proved to be a difficult task for some robbers, especially when they
were armed with knives because this weapon seems less suitable for making
targeted expanded body postures. Robbers who moved rapidly into copre-
sence with the victims as they came running in with weapons ready, mostly
guns in this tactic, faced a different challenge. They had to give the victims
22 Journal of Research in Crime and Delinquency 55(1)

time to adapt to the robbery frame they were suddenly proposing. Guns,
rather than knives, seemed to fit more easily with aggrandizing posturing,
thus offering more support to robbers’ claim to dominance, in addition to
their more lethal power. In the cases where resistance occurred, robbers
either did not succeed in attaining emotional dominance, as expressed by
their insecure movements or minimizing postures and related weapon
usage, or they failed to impose a robbery frame on puzzled victims who
did not seem to comprehend the situation. We conclude that the forceful-
ness of armed threats to establish a robbery frame is related to the types of
weapons used and to what extent these weapons afford aggrandizing body
posturing. In order to explain why some robbers with a gun fail and some
with less lethal weapons like hammers, knives, screwdrivers, or even bare
hands succeed, the role of weapons should thus be seen as part of the whole
array of bodily strategies to impose the robbery frame.
Perhaps the most important conclusion is that video analysis opens up a
new, close-up microperspective on the interactions that make up violent
crime: One that perceives them as sequences of body movements. Future
studies that want to follow this path might benefit from the following
methodological notes. First, the density and complexity of video data seem
to be the prime obstacle for analyzing it properly. We propose four sugges-
tions to make sense of such a wealth of information: (a) become familiar
with the data by watching it multiple times to get a sense of the general flow
of interactions; (b) work in a team, compare independently, made observa-
tions, and organize joint video sessions; (c) clearly specify what specific
modes of behavior are to be analyzed based on the literature and initial
observations; (d) when coding the data, focus, with each run-through, on a
specific mode of behavior. Second, one disadvantage of the video footage
used here is that they lack audio. Audio could have been an important
additional indicator for observing the emotional signaling and impact of
certain body movements and posturing. Third, while we did not pay atten-
tion to the use of physical space by robbers and victims, this seems a
relevant issue to consider in more detail. Video analysis of how robbers
and victims use entrances, exits, counters, display shelves, and other objects
can reveal how the physical setup influences the sequences of body move-
ments that make up violent crime. Finally, we recommend that future stud-
ies complement video data with written information of various sorts on
what happened (e.g., judicial case files, news reports, or interviews). This
is not only useful to verify the observations but also to gain a better under-
standing of all body movements that make up the situation as an integral
hermeneutic unit.
Mosselman et al. 23

We end with an ambitious suggestion for future studies. Given the con-
tradicting results of prior studies of weapon use by Dutch and U.S. research-
ers, we think it is worthwhile to analyze weapon use as part of the bodily
dynamics in comparative, larger scale video analyses of robberies in various
societies and settings. Such comparative project should involve both qua-
litative, interpretative analyses and the fine-grained coding that allows to
statistically assess (preferably sequence analyses given the real-time data)
the relationships between various weapons and forms of body posturing of
both victims and offenders.

Appendix

Code Description Interpretation

Weapon Weapon visible for other actors Sign of

visible dominance
Weapon Weapon is directly able to injure: pointed gun/raised Sign of
threat weapon/ready to throw or stab dominance
Weapon Weapon is used with the intention to injure/ Sign of
use dominate/force the other party dominance
Physical Touch/hit/push/grab Sign of
force dominance
grab
Minimize Body language: avoiding gaze (head down/bowed), Sign of
posture postures and movements that are shrinking, head submission
ducked down, hands cover face, chest inward, and
shoulder slumped forward. Tries to hide
Maximize Body language: strong physical presence by making Sign of
posture oneself large, for example, by standing erect and/or dominance
with hands on hips. Bodily openness. Pride; small
smile, head tilted slightly (approximately 20
degrees) back, and expanded posture
Hands up Hand(s) up or in front of chest; open hands. Open Sign of
hand toward another actor, sign of self-protection submission
Move away Moves back or give ground to another actor. Sign of
submission
Move Move face-front toward another actor Sign of
toward dominance
Compliance Acts consciously in accord with the other actors’ will Sign of
(give values, lay on the floor) submission
Resistance Acts consciously against the other actors’ will (give Sign of
values, lay on the floor) dominance
24 Journal of Research in Crime and Delinquency 55(1)

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research,
authorship, and/or publication of this article.

Funding
The author(s) received no financial support for the research, authorship, and/or
publication of this article.

Note
1. We thank the anonymous reviewer for bringing this distinction to our attention.

References
Bernasco, W., M. Lindegaard, and S. Jacques. 2013. “Overvallen Vanuit Daderper-
spectief: Situationele Aspecten Van Gewelddadige, Niet-Gewelddadige En
Afgeblazen Overvallen.” Politiewetenschap, 69. Apeldoorn, the Netherlands:
Politie & Wetenschap.
Collins, R. 2008. Violence: A Micro-sociological Theory. Westport, CT: Green-
wood Publishing Group.
Cook, P. 1981. The Role of Firearms in Violent Crime: An Interpretive Review of the
Literature, with Some New Findings and Suggestions for Future Research. Dur-
ham, NC: Institute of Policy Sciences and Public Affairs, Duke University.
Copes, H., A. Hochstetler, and M. Cherbonneau. 2012. “Getting the Upper Hand:
Scripts for Managing Victim Resistance in Carjackings.” Journal of Research in
Crime and Delinquency 49:249-68.
Cowan, K. 2014. “Multimodal Transcription of Video: Examining Interaction in
Early Years Classrooms.” Classroom Discourse 5:6-21.
Ekman, P. 1991. Telling Lies. New York: Norton.
Feeney, F. 1986. “Robbers as Decision Makers.” Pp. 53-71 in The Reasoning Crim-
inal: Rational Choice Perspectives on Offending, Research in Criminology,
edited by D. B. Cornish and R. V. Clarke. New York: Springer.
Felson, R. B. 1982. “Impression Management and the Escalation of Aggression and
Violence.” Social Psychology Quarterly 45:245-54.
Felson, R. B. and H. J. Steadman. 1983. “Situational Factors in Disputes Leading to
Criminal Violence.” Criminology 21:59-74.
Felson, R. B. and J. T. Tedeschi. 1993. Aggression and Violence: Social Interac-
tionist Perspectives. Washington, DC: American Psychological Association.
Goffman, E. 1959. The Presentation of Self in Everyday Life. London, UK: Penguin
Books.
Mosselman et al. 25

Hall, J. A., E. J. Coats, and L. S. LeBeau. 2005. “Nonverbal Behavior and the
Vertical Dimension of Social Relations: A Meta-analysis.” Psychological Bul-
letin 131:898-924.
Heidegger, M. 1927/2002. Being and Time. New York: Harper (first published in
1927 as Sein und Zeit).
Jacobs, B. A. 2012. “Carjacking and Copresence.” Journal of Research in Crime
and Delinquency 49:471-88.
Kleck, G. 2005. Point Blank: Guns and Violence in America. New Brunswick, NJ:
Transaction.
Kleck, G. and M. A. DeLone. 1993. “Victim Resistance and Robber Weapon Effects
in Robbery.” Journal of Quantitative Criminology 9:55-81.
Kleck, G. and K. McElrath. 1991. “The Effects of Weaponry on Human Violence.”
Social Forces 3:669-92.
Klusemann, S. 2010. “Micro-situational Antecedents of Violent Atrocity.” Socio-
logical Forum 25:272-95.
Knoblauch, H. 2009. “Social Constructivism and the Three Levels of Video Anal-
ysis.” Pp. 1-198 in Video Interaction Analysis. Methods and Methodology, edited
by T. U. Kissmann. Frankfurt am Main: Peter Lang GmbH. Internationaler
Verlag der Vissenschaften.
Lindegaard, M. R., W. Bernasco, S. Jacques, and B. Zevenbergen. 2013. “Posterior
Gains and Immediate Pains: Offender Emotions Before, During and After Rob-
beries.” Pp. 58-76 in Affect and Cognition in Criminal Decision Making:
Between Rational Choices and Lapses of Self-control, edited by J. van Gelder,
H. Elffers, D. Reynald, and D. Nagin. New York: Routledge.
Lindegaard, M. R., W. Bernasco, and T. de Vries. 2016. Overvallen in Beeld: Dader,
Slachtoffer en Omstander Gedrag. Den Haag/Apeldoorn: Elsevier-Reed/Politie
& Wetenschap.
Luckenbill, D. F. 1981. “Generating Compliance: The Case of Robbery.” Urban
Life 10:25-46.
Luckmann, T. 2006. “Some Remarks on Scores in Multimodal Sequential Analysis.”
Pp. 29-34, in Video Analysis: Methodology and Methods, edited by H. Knoblauch,
B. Schnettler, J. Raab, and H. G. e Soeffmer (a cura di). Lang: Frankfurt am
Main.
Overvallen, NL. 2016. Per Sector. Politie. Rapport gegenereerd op 8 maart 2016.
[Police report about robberies in the Netherlands]. https://www.politie.nl/bin
aries/content/assets/politie/documenten-algemeen/publicaties-archief/2012-
gewelddadige_vermogenscriminaliteit_-overvallen-en-straatroof.pdf.
Polk, K. 1999. “Males and Honor Contest Violence.” Homicide Studies 3:6-29.
Rovers, B., M. Bruinsma, M. Jacobs, M. Jans, H. Moors, M. Siesling, and C. Fijnaut.
2010. Overvallen in Nederland; Een fenomeenanalyse en evaluatie van de
26 Journal of Research in Crime and Delinquency 55(1)

aanpak. [Robberies in the Netherlands. Analysis and evaluation]. Den Haag, the
Netherlands: Boom Juridische Uitgevers.
Scheff, T. J. and S. M. Retzinger. 1991. Emotions and Violence: Shame and Rage in
Destructive Conflicts. Lexington, KY: Lexington Books.
Siegel, S. M., M. L. Friedlander, and L. Heatherington. 1992. “Nonverbal Relational
Control in Family Communication.” Journal of Nonverbal Behavior 16:117-39.
Wells, W. and J. Horney. 2002. “Weapon Effects and Individual Intent to Do Harm:
Influences on the Escalation of Violence.” Criminology 40:265-96.
Wilkinson, D. L. and J. Fagan. 2001. “A Theory of Violent Events.” Pp. 169-95 in The
Process and Structure of Crime: Criminal Events and Crime Analyses, edited by
R. F. Meier, L. W. Kennedy, and V. F. Sacco. New Brunswick, NJ: Transaction.
Wright, R. T. and S. H. Decker. 1997. Armed Robbers in Action: Stickups and Street
Culture. Boston, MA: Northeastern University Press.

Author Biographies
Floris Mosselman is research assistant at the Group Violence Research Programme,
Department of Sociology at the University of Amsterdam, and a junior lecturer at
the Erasmus University Rotterdam. His research interests are the interactionist
aspects of deviant, mostly violent, behavior, video-analysis and qualitative methods.
Don Weenink is a member of the Department of Sociology at the University of
Amsterdam. He currently supervises the Group Violence Research Programme. His
main research interests are violence and emotions in social life.
Marie Rosenkrantz Lindegaard is a senior researcher at the Netherlands Institute
for the Study of Crime and Law Enforcement (NSCR) and an associate professor at
the department of Sociology of the University of Copenhagen. Her research interests
are situational aspects of crime, agency, street culture, qualitative methods, use of
camera footage for crime research, and urban ethnography in South Africa.